We carefully design the potential functions to enable relational reasoning among different objects according to knowledge graph embedding methods.

While this has been demonstrated empirically overtimeinavarietyoftasks,theoretical resultsquantifying thebenefits of such methods are still few and far between. In this paper, we discuss the kind of structure in a Markov decision process which gives rise to efficient HRLmethods.

Machine learning (ML) algorithms can solve many prediction tasks with greater accuracy than classical methods.

We additionally assume that the agent cannot interact with the environment but has access to the action-labeled transition data collected by some agents with unknown qualities.