Now that we have solved the GridWorld problem, there are other practical aspects in reinforcement learning and overall deep learning phenomena that need to be considered too. In this section, we will see some frequently asked questions that may be already on your mind. Answers to these questions can be found in Appendix.
- What is Q in Q-learning?
- I understand that we performed the training on GPU and cuDNN for faster convergence. However, there is no GPU on my machine. What can I do?
- There is no visualization, so it is difficult to follow the moves made by the agent toward the target.
- Give a few more examples of reinforcement learning.
- How do I reconcile the results obtained for our mini-batch processing? ...