In this chapter, we covered an emerging technique in RL called Imitation Learning or Behavioral Cloning. This technique, as we learned, takes the captured observations of a player playing a game and then uses those observations in an online or offline setting to further train the agent. We further learned that IL is just a form of Transfer Learning. We then covered a technique with ML-Agents that will allow you to transfer brains across environments. Finally, we looked at how to chain IL andtransfer learningas a way of stimulating the agent's training into developing new strategies on its own.
In the next chapter, we will further our understanding of DRL in games by looking at multiple agent training scenarios.