Learning personalized reward functions with Interaction-Grounded Learning (IGL)

Rewards play a crucial role in reinforcement learning (RL). A good choice of reward function motivates an agent to explore and learn which actions are valuable. The feedback that an agent receives via rewards allows them to update their behavior and learn useful policies. However, designing reward functions is complicated and cumbersome, even for domain experts. Automatically inferring a reward function is more desirable for end-users interacting with a system.