Inverse reinforcement learning

Concept

A method used in robotics to infer a reward function from observed human behavior, assuming humans act optimally with respect to their preferences.

Mentioned in 2 videos