Generalized Beliefs for Cooperative AI