Inreinforcement learning, option models (Sutton, Precup & Singh, 1999; Precup, 2000) provide the framework for this kind of temporally abstract prediction and reasoning. Natural intelligent agents are also able to focus their attention on courses of action that are relevant or feasible in agiven situation, sometimes termed affordable actions.

Offline reinforcement learning (RL) aims to learn policies from pre-existing datasets without further interactions, making it a challenging task. Q-learning algorithms struggle withextrapolation errors inofflinesettings, while supervised learning methods are constrained by model expressiveness.

Deceptive agents are a challenge for the safety, trustworthiness, and cooperation of AI systems.

In the above example, the state is which candidate will make more progressonnon-partisanissues.

In the above example, the state is which candidate will make more progressonnon-partisanissues.