Reinforcement learning Reinforcement learning follows a different strategy where the output of the system is not a single action; it’s a sequence of actions. So the single action is not important here; what is important is the policy, that is, the sequence of correct actions to reach the goal. So, we have to formulate a policy of sequence of actions which leads to the right, correct destination. For example, you start at your home, and there are multiple ways to reach a certain place which is new to you. So, you started moving in one direction, and if you could be able to reach your destination with the lesser shortest length or without any hurdles or block of routes, like that, that route can be recorded in your map because in the future, you may prefer the route, right? So similarly, you can take another route and find out whether it may take you to the destination. A very simple example, a very common example, of what reinforcement learning can be compared as, is chess playing. So, in a chess-playing game, you know, there are two players who started playing with the preferred color of coins. They started moving, and winning is not decided or any winning moment milestones is not decided by a single action; it’s a sequence. The reinforcement learning has two components: one is the agent, another one is the environment. The agent is allowed to perform certain action Y onto the environment. If the action is a successful action, then the agent’s state is changed from SI to S_t+1, and it is rewarded. Otherwise, the agent is penalized.