Question $1$

$($a$)$ Describe in layperson$$s terms what the Deepmind DQN agent does. $[4]$

$($b$)$ Describe the raw data that is fed to the DQN agent. $[4]$

$($c$)$ How is the raw data processed before it is fed to the neural network? $[4]$

$($d$)$ For each process you described in the previous question, state why that

process is applied. $[6]$

$($e$)$ Explain how the training works in DQN$.$ What is the input to the network?

What is the output? What is the error metric. $[6]$

$($f$)$ Compare training in DQN to a more typical training process where the

input data is a fixed set. $[6]$