Reinforcement learning
AIRobots that make judgments like humans are coming faster than we think
KAIST researchers built VOTP, a system that learns human judgment from a few videos to train robots more efficiently.
Joseph Shavit