Imitation Learning
A machine learning approach where the model learns to perform a task by observing and mimicking expert demonstrations rather than through explicit reward signals.
Methods
Behavioral cloning: Directly learn a mapping from states to expert actions (supervised learning). Inverse RL: Infer the reward function the expert is optimizing. DAgger: Iteratively collect expert feedback on the learner's mistakes.
Applications
Robotics (learning manipulation from human demos), autonomous driving (learning from human driving data), game AI, and LLM alignment (RLHF starts with imitation of human-preferred responses).