World Models

Can agents learn inside of their own dreams?