Google Deepmind DQN plays Atari Breakout | NickysChannel13

Google Deepminds deep learning algorithm DQN learns to engage in Atari Breakout. It designed some noticeable progress and even located out, that it is most efficient to dig a tunnel and enable the ball run at the rear of the wall.
The algorithm knew very little about the controls. It was fed with raw pixel details and had to make a decision on the next motion. It had no plan of the theory of a ball and the activity itself. The only intention was to optimize the rating.