Episode Details

Proposes a variational neural dynamics framework for continual robot policy learning, enabling robots to acquire new skills without forgetting previously learned ones.