Summary: Play Video Games with Generative AI

In this chapter, we explored another application of generative models in reinforcement learning. First, we described how RL allows us to learn the behavior of an agent in an environment and how deep neural networks allowed Q-learning to scale to complex environments with extremely large observation and action spaces.

We then discussed inverse reinforcement learning and how it varies from RL by “inverting” the problem and attempting to “learn by example.” We discussed how the challenge of comparing a proposed network to an expert network can be evaluated using entropy. We explored how GAIL is just one of many possible formulations of this general approach, each using different loss functions. Finally, we implemented GAIL using the bullet-gym physics simulator and OpenAI Gym.

In the final chapter, we’ll conclude by exploring recent research in generative AI in diverse problem domains, including bioinformatics and fluid mechanics. We’ll also provide references for further reading as you continue to discover this developing field.