Action-Adaptive Continual Learning: Enabling Policy Generalization under Dynamic Action Spaces

Open in new window