Understanding Reinforcement Learning As Fast As Possible

Reinforcement Learning is a type of ML Algorithm. The Agent in a Reinforcement Learning takes actions on the environment and interacts, in order to receive a reward, to make a sequence of decisions. The reward can be positive or negative, called penalties. The aim of the agent in RL is to gain maximum rewards. Unlike supervised learning, the agent does not know what actions it has to take in order to get a reward. The agent learns this by itself by performing trial-error. The designer has to set what is known as the Reward policy, i.e. the rules with no hints or suggestions. The RL models can outperform humans by giving them enough time to improve themselves using a sufficiently powerful infrastructure.

RL is seen in Industrial Automation, Games, ML and other applications. When it comes to Healthcare RL is seen in decision making, optimal drug delivery personalized for each patient, for HIV clinical Trials, reduction in epilepsy and so on.