# What is Reinforcement Learning?

It is the process of artificial intelligence learning to make correct decisions through trial and error and by using a reward mechanism.

## Overview
In this method, the artificial intelligence is not directly told what to do. Instead, it moves within an environment and receives either a reward (points) or a penalty for every action it takes. Over time, it develops the strategy to maximize its reward on its own.

*Analogy: It is like teaching a dog to sit by giving it a treat when it sits and not giving one when it doesn't; the dog learns what it needs to do to get the reward.*

## How it works
The artificial intelligence performs trials thousands of times within a simulation. While increasing the value of successful moves, it eliminates the erroneous ones to find the path that yields the best result.

## Where it is used
It is used in game-playing artificial intelligences, robotic motion control, and financial trading strategies.

## Commonly confused with
It can be confused with supervised learning; in that, the correct answer is provided, whereas in this, only the result is rewarded.

## Frequently asked questions
**Why does it take so much time?**
Because the system needs to perform many trials and learn a lesson from each trial in order to find the correct path.


## Related terms
- [RLHF](/en/dictionary/rlhf/)
- [Deep Learning](/en/dictionary/deep-learning/)
- [Machine Learning](/en/dictionary/machine-learning/)

## Related tools
- [Prime Agent](/en/discover/prime-agent/)
- [Miles](/en/discover/miles/)
- [Microduck Rl](/en/discover/microduck-rl/)

---
Source: TreScout Dictionary · https://trescout.com/en/dictionary/reinforcement-learning/