Generative Adversarial Imitation Learning

Sanket Gujar
8 min readApr 21, 2018

Learning from demonstrations will play a very important role in the age of robotics. If the robots or humans need to survive with each other, robots need to predict and understand human behaviors for reliability and safety.

The interest here is (you guessed it) imitation learning which is learning to perform a task from demonstrations. For a Reinforcement Learning agent to do well they need…

--

--

Sanket Gujar

Computer Science Graduate Student at WPI, Former Perception Intern at Luminar tech, PA. sanketgujar.github.io