This is a step forward in the field of artificial intelligence made by the company DeepMind which created an algorithm capable to expand its knowledge by learning.

Artificial intelligence called DQN (Deep-Q-Network), learns to play games such as "Space Invaders" until it manages to beat a human player. Mostly it is about games that were popular on the Atari platform.

It learns the gaming system directly from experience in order to manage a wide range of difficult tasks.

So, DQN learns while working and accumulating experience. For example, it learned to play the games by randomly pressing buttons and registering points when its choice of a particular button brought a good result.

It was also stated that this kind of artificial intelligence in the future could prove usefull in assisting in areas such as testing weather conditions or diseases that involve numerous and very complex data.