Papers with Code paper Jun 17

Playful Agentic Robot Learning

Current agentic robot systems can write executable Code-as-Policy programs, observe feedback, and revise behavior across multiple attempts, but they remain largely task-driven: reu...