July 2018
Intermediate to advanced
254 pages
8h 12m
English
In the previous chapter, we discussed the
-greedy action selection policy which takes the best action as per the agent's action-value estimate with a probability of 1-
and takes a random action with a probability given by epsilon,
. Epsilon is a hyperparameter that can be tuned based on the experiments to a good value. A higher value of means that the agent's actions will be random and a lower value ...
Read now
Unlock full access