Hands-On Reinforcement Learning with Python

Chapter 6

1. An MAB is actually a slot machine, a gambling game played in a casino where you pull the arm (lever) and get a payout (reward) based on a randomly generated probability distribution. A single slot machine is called a one-armed bandit and, when there are multiple slot machines it is called multi-armed bandits or k-armed bandits.
2. An explore-exploit dilemma arises when the agent is not sure whether to explore new actions or exploit the best action using the previous experience.
3. The epsilon is used to for deciding whether the agent should explore or exploit actions with 1-epsilon we choose best action and with epsilon we explore new action.
4. We can solve explore-exploit dilemma using a various algorithm such epsilon-greedy policy, softmax ...