CoRL 2026 workshop makes the case for continually self-improving robots
chris_j_paxton · x · 2026-11-12
The Continually Self-Improving Robots workshop at CoRL 2026 (Nov 12, Austin) argues that imitation alone can't close the gap between broad initial competence and reliable real-world performance — robots must autonomously collect data, discover failures, and refine skills through interaction, an "era of experiences" for robot learning.
Highlights:
- Invited speakers: Abhishek Gupta (UW & Amazon FAR), Kay Ke (Physical Intelligence), Jeff Clune (UBC & Recursive), Jason Ma (Dyna), Pete Florence (Generalist)
- Spotlight talks include test-time gradient guidance for flow policies and WEAVER, a world model for robotic manipulation
- Accepted papers are public on OpenReview, with paper awards and breakout sessions
More from Embodied
- Local AI Daily: DeepSeek-V4.1-Flash hits 90 tok/s on 4× DGX Spark; Runway open-sources Praxis-1 — AIFlow_ML · 2026-11-12
- CoRL goes out Texas-style: mechanical bull riding night, cowboy hats TBD — KyleMorgenstein · 2026-11-12
- Tesla Robotaxi Now Runs 24/7 Across 8 US Cities — XFreeze · 2026-11-12
- RobotEra humanoids sort parcels at SF Express hub during Double 11, 1,200 parcels per hour — CyberRobooo · 2026-11-12
- Hackathon Project Uses Hermes Agent to Turn Prompts into 3D-Printable CAD Designs — alexcovo_eth · 2026-11-12
- Direct-drive liquid-cooled robot hand debuts, targeting dexterity at the actuator level — ybisk · 2026-11-12