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.
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
Related tools
This explanation was written in plain language for TreScout and machine-translated from the Turkish original · the Turkish version prevails. If something looks wrong or missing, write to hello@trescout.com. Read in Turkish →