Canonical event

Inner Monologue closes language-model planning loops with environment feedback

Researchers showed language-model planners incorporating success detection, scene descriptions and human interaction to replan in simulated and real robot tasks.