Use the exercises in this section to enhance and reinforce your learning. Attempt at least a few of these exercises on your own, and remember this is really for your benefit:
- Set up and run the 3DBall example environment to train a working agent. This environment uses multiple games/agents to train.
- Set the 3DBall example to let half of the games use an already trained brain and the other to use training or external learning.
- Train the PushBlock environment agents using external learning.
- Train the VisualPushBlock environment. Note how this example uses a visual camera to capture the environment state.
- Run the Hallway scene as a player and then train the scene using an external learning brain.
- Run the VisualHallway scene as a player and then train the scene using an external learning brain.
- Run the WallJump scene and then run it under training conditions. This example uses Curriculum Training, which we will look at further in Chapter 9, Rewards and Reinforcement Learning.
- Run the Pyramids scene and then set it up for training.
- Run the VisualPyramids scene and set it up for training.
- Run the Bouncer scene and set it up for training.
While you don't have to run all these exercises/examples, it can be helpful to familiarize yourself with them. They can often be the basis for creating new environments, as we will see in the next chapter.