Introduction to Reinforcement Learning

WHAT IS MACHINE LEARNING?

Machine Learning is a subset of Artificial intelligence, that deals with making computers learn from past experiences with the help of algorithms. However, the machine need not be explicitly programmed, and it grasps knowledge from its own experiences. 

Basically, Machine learning is classified into 4 different parts:

• Supervised Learning
• Unsupervised Learning
• Semi-Supervised Learning
• Reinforcement Learning

Here, in this series, we are going to discuss the concept of Reinforcement Learning in detail. 

WHAT IS REINFORCEMENT LEARNING?

1. Reinforcement Learning is a type of Machine Learning method where the AI agent is trained to learn with time.
2. AI agent is developed in such a way that, it iteratively(repeatedly in easy terms) interacts with the environment and performs some specific actions.
3. For every action, feedback is given to the AI agent. The feedback is in such a form, that the agent gets a positive reward for every positive action and it gets a penalty as feedback for every negative action. 
4. The AI agents aim to perform those actions that lead to getting the maximum of the rewards(positive feedback) as an output. However, learning from the actions, that results in negative feedback, thus avoiding them.
5. Note that the negative feedback that is provided to the agent results in a change of the state of the agent. Also, the AI agent will remain in the same state until he continuously receives positive feedback. However, you can say that the agent need not be explicitly programmed to perform a specific kind of action.
6. It learns with its own feedback and experiences. Since it performs action according to the environment, we can conclude that it does not work on labeled data like Supervised learning.