- What is the difference between using Keras as a wrapper for TensorFlow and using TensorFlow by itself?
- What kind of prediction problem is CartPole? What type of target is being predicted?
- What advantage does binning the state space provide?
- Give a 2–3 sentence explaining how experience replay works.
- Explain the moving target problem in estimating Q-values.
- Explain what maximization bias is.
- When does maximization bias tend to happen?
- What advantage do dueling DQNs provide over regular DQNs?




















































